Abstract
Background
In 2001, an international beverage company implemented an HIV workplace programme providing free antiretroviral treatment (ART) for employees and dependents in sub-Saharan Africa, at a time when ART, cost assessments of ART programmes and related public funding was hardly available. This study describes the outcomes of this programme with respect to achieving the UNAIDS 90-90-90 targets in five African countries and analyses trends over the past 15 years.
Methods
Anonymous human resource data were analysed in three cohorts of participants (those enrolling in 2001– 2005, 2006–2010 and 2011–2015).
Results
Over 15 years, 42,490 unique individuals in five African countries were tested for HIV in this programme and 746 (1.8%) were found to be HIV-infected. Between 2002 and 2015, the proportion of HIV-positive participants on ART increased from 42% to 94% and the proportion of participants on ART who achieved virological suppression increased from 38% to 87%.
Conclusions
This study shows that in one of the earliest HIV treatment programmes in Africa long-term success has been achieved, approaching the current UNAIDS 90-90-90 targets, demonstrating that the treatment of HIV in developing countries is possible with superior results at low costs (45 US dollars/employee). Reasons for this success include continuous access to on-site quality care and ART and the assistance of an independent NGO with experience in HIV treatment. This provides an argument to continue private sector involvement in international efforts to combat HIV/AIDS, particularly in light of increased ART targets, under-capacity in the public sector and stagnating international funding.
Introduction
Over the past 15 years, access to antiretroviral treatment (ART) for people living with HIV worldwide has increased dramatically and by the end of 2017, more than 21 million people were on ART [1]. This is in contrast to the beginning of this millennium, when an estimated 21 million people were living with HIV in sub-Saharan Africa, with only 11,000 having access to ART (representing a coverage of 0.05%) [2]. In resource-rich settings, ART had been available since 1996, but due to the high costs and (perceived) logistical and other barriers of ART in resource-limited settings, access in sub-Saharan Africa remained limited.
In 2001, an international beverage company decided to implement one of the very first HIV treatment workplace programmes in its operating companies in Africa, as well as in some countries in Asia. These programmes provided extensive HIV prevention measures as well as free HIV testing, counselling and ART for HIV-infected employees and their dependents. The Joep Lange HIV Workplace Program in Africa targeted staff and dependents in Rwanda, Burundi, Democratic Republic of the Congo (DRC), Republic of the Congo, Nigeria and Sierra Leone. Medical expertise and technical assistance to this programme was provided by PharmAccess Foundation, a non-governmental organization based in the Netherlands, under leadership of late Professor Joep Lange. PharmAccess provided training to clinicians and lab technicians, assistance with procurement and importation of antiretroviral drugs and installed essential laboratory equipment. Moreover, PharmAccess provided quality support by establishing clinical databases for optimal patient management, performing (interim) data quality checks and analyses, monitoring and evaluation and providing telephone feedback and expert advice to the medical staff. Importantly, PharmAccess was checking on behalf of the company whether programme funds were indeed spent on the agreed HIV treatment and care packages.
To date, the Joep Lange HIV Workplace Program has been running in 23 clinics over the past 15 years. An evaluation of the first 5 years of the programme has previously been described [3]. This study analyses the 15-year outcomes of the workplace programme in order to evaluate the potential benefits of a private sector HIV programme in sub-Saharan Africa. In addition, we estimate the programme's progress towards the two treatment-related UNAIDS 90-90-90 targets: the proportion of participants on ART and the proportion of participants on ART reaching virological suppression
Methods
Programme
The Joep Lange HIV Workplace Program started in 2001 in five African countries between 2001 and 2004, as has previously been described in detail [3]. The programme includes a comprehensive HIV prevention package (for example, condom distribution, STD treatment and prevention of mother-to-child transmission), voluntary HIV testing and free ART for employees, their spouses and children, as well as for ex-employees. Once on ART, participants are entitled to receive lifelong HIV care, even if the employee is laid off from work at the company. HIV care is provided by the local company doctor and nurses, and all HIV test results remain confidential through separated data systems. Company management has no access to the test results and HIV testing is not part of the selection procedure for new employees. HIV testing in the different Operating Companies is either offered as voluntary counselling and testing or as an opt-out procedure, for example, during yearly health check-ups. Awareness and counselling and testing events are organized at the company's premises on World AIDS Day every year. HIV testing is done using two different rapid tests according to the national guidelines of the pertinent countries. For children under 18 months of age, RNA PCR testing is done.
Participants were enrolled in the programme when tested HIV-positive and after informed consent. Enrolled participants did not automatically start ART, but only when they met the eligibility criteria for treatment according to guidelines of the European AIDS Clinical Society (EACS) at that time [4]. Viral load and CD4+ T-cell count monitoring was conducted every 6 months. HIV drug resistance testing was available for patients whose plasma viral load rebounded despite being considered adherent to their treatment. First-line treatment was either efavirenz (EFV)- or nevirapine (NVP)-based. Women who were (planning to get) pregnant received protease inhibitor (PI)-based ART. Participants with treatment failure or with documented resistance against EFV or NVP were switched to a PI-based second-line regimen.
Data management
All patients signed an informed consent form, giving permission for data collection and analysis of anonymized data. Data were kept in a centrally located database (Lotus Notes®; IBM, Armonk, NY, USA). Data were entered by the physician or nurse at each site, using a unique patient identifier number. Names of patients were encrypted, precluding access to data by the human resource department or any user other than the treating physician or nurse. An anonymized database was accessible to medical experts of the PharmAccess network in Amsterdam, the Netherlands, who evaluated the quality of care and provided guidance on individual patient management through monthly teleconferences with the physicians and nurses on site. This provided an additional check that the funds were spent on the agreed terms of the clinical protocol.
Data analysis
This analysis is restricted to the five sub-Saharan Africa countries, which had at least 10 participants enrolled in the programme: Nigeria, DRC, Republic of the Congo, Burundi and Rwanda. 15-year follow-up data from 2001 to 2015 were analysed. Participants’ characteristics were summarized separately for those entering the programme (not necessarily starting ART) between 2001 and 2005 (cohort 1), 2006 and 2010 (cohort 2) and 2011 and 2015 (cohort 3). Data for participants in all three cohorts were censored on 31 December 2015. Duration of follow-up, therefore, differed between cohorts and was longest for participants in cohort 1 and shortest for participants in cohort 3. Participants in the programme were either employees of the company, spouses of employees, children of employees or ex-employees and their dependents.
To assess the workplace programme results regarding the treatment-related UNAIDS 90-90-90 targets (% on treatment, % with virological suppression), we used the following data: to analyse the proportion of HIV-infected participants on treatment per calendar year (2001-2015), we used data from a participant's last visit of each calendar year to calculate the number of participants on treatment divided by the number of participants enrolled in the programme; to evaluate the number of participants on ART with virological suppression, we used a participant's last available viral load (VL) of each calendar year (2001-2015). The denominator for this calculation was the number of participants on ART. The VL assays used in the first years of the programme had a lower limit of detection of <400 copies/ml, which was, therefore, used as a definition of virological suppression for comparisons over the full 15-year programme period. Participants without any VL results in a particular year were not included in the analysis of that year (as-observed analysis). Comparisons between the three cohorts were conducted using a χ 2 or Kruskal–Wallis test. A two-sided P-value of ≤0.05 was considered significant. Data were analysed using Stata 12® (StataCorp LP, College Station, TX, USA).
Results
A flow chart of enrolment and follow-up in this study is presented in Figure 1.

Flow chart of study enrolment and follow-up (2001–2015)
HIV Testing
From September 2001 to December 2015, 42,490 unique individuals (employees, spouses, children and ex-employees) were tested for HIV-1 as part of the workplace programme. Of these, 746 (1.8%) were found to be HIV-infected and were enrolled in the workplace programme in 1 of 23 sites in Nigeria (11), DRC (6), Republic of the Congo (2), Rwanda (2) and Burundi (2; Table 1). During the programme 5,141 individuals had more than one HIV test done (range 2-16), resulting in the identification of 31 (0.6%) sero-converters.
Characteristics of 702 participants at enrolment in the Joep Lange HIV Workplace Program, by cohort
Cohort 1: first enrolled between 2001 and 2005; cohort 2: first enrolled between 2006 and 2010; cohort 3: first enrolled between 2011 and 2015. CDC, Center for Disease Control and Prevention; DRC, Democratic Republic of the Congo.
In cohort 1 (2001-2005), 7,053 individuals had a first HIV test in the programme of whom 431 were found to be positive (6.1%), in cohort 2 (2006-2010), 174 of 8,701 (2.0%) first-time testers were positive, and in cohort 3 (2011-2015) 110 out of 26,730 (0.4%), P<0.001. The median CD4+ T-cell count at diagnosis increased from 271 (IQR 140-432) cells/mm3 in cohort 1, and 280 (IQR 181-449) cells/mm3 in cohort 2, to 340 (IQR 215-569) cells/mm3 in cohort 3, P=0.0034.
Antiretroviral treatment
In 2001, 52.5% of participants in the programme had started ART at their last clinic visit of that calendar year and this proportion increased to 83.7% in 2010, and 85.7-93.9% in the most recent 3 years (Figure 2). In 2001, the 23 participants who started ART were all on a first-line non-nucleoside reverse transcriptase inhibitor based regimen (100%). By 2005, 24/246 (9.8%) regimens were PI-based, and in 2015, the number of participants on PI-based ART had increased to 59/438 (13.5%). Both the use of stavudine (d4T) and didanosine (ddI) has been phased out during the course of the programme. After 2011, none of the participants were on d4T and <3% was on ddI (Figure 3). In 2001, the most frequently used regimen was zidovudine (AZT)/lamivudine (3TC)/EFV (76.0%), in 2005 AZT/3TC/NVP (47.2%), and in 2010 and 2015 tenofovir disoproxil fumarate (TDF)/emtricitabine (FTC)/EFV (37.0% and 51.8%, respectively).

Number of participants enrolled, on treatment and with virological suppression per year

Antiretroviral treatment regimens used in the workplace programme over time
Virological suppression
Figure 4 shows the proportion of participants achieving virological suppression, 6 to 36 months after initiating ART for the three cohorts separately. In cohort 1, 48.4–68.5%, in cohort 2, 59.3–86.8%, and in cohort 3, 72.4–93.8% achieved virological suppression after 6 to 36 months of ART. Overall, the proportion with virological suppression per year increased from 37.5% in 2002, to 80.6% in 2010 and 86.6–92.0% in 2013– 2015 (Figure 2).

Proportion of participants with virological suppression, 6 to 36 months of antiretroviral treatment initiation, by cohort
Mortality
Over 15 years of follow-up, 99 HIV-infected participants died (14.1%): 55 employees (15.3%), 18 (7.8%) spouses, 19 (26.4%) children and 7 (18.0%) retired employees. The cause of death was deemed HIV-related in 18 (18.2%) cases (mainly opportunistic infections in participants who, for any reason, stopped ART), not HIV-related in 42 cases (42.4%), and was unclear or not documented in the remaining 39 (39.4%) cases. 36 (36.0%) participants died within 1 year after enrolment: 21/408 (5.2%) in cohort 1, 9/172 (5.2%) in cohort 2 and 6/122(4.9%) in cohort 3, P=0.992.
Discussion
In 2001, an international beverage company started one of the first HIV/AIDS workplace programmes in Africa, in close collaboration with PharmAccess. The company decided to pay for the treatment of its infected employees and their dependents at a moment when public funding was not available and the costs to the company were, prior to the launch of this programme, largely unknown. A previous analysis of this programme showed that the costs of providing ART to the company's employees in the first 5 years of the programme was around 45 US dollars per employee per year [5]. This manuscript evaluates this large HIV workplace programme in five countries in sub-Saharan Africa over the past 15 years. In the most recent years up to 94% of participants were on ART, and up to 92% had virological suppression, reaching the treatment-related UNAIDS 90-90-90 targets.
The proportion of individuals who tested positive for HIV decreased significantly from 7.9% in the first years of the programme to 0.4% in 2011–2015 which is most likely due to a change in demographics of the population tested (selection bias). At the beginning of the programme, individuals who were most at risk or who already had symptoms of HIV infection were reporting for testing, while in later stages also the people without symptoms and with less risk factors were enrolled. This is exemplified by the median CD4+ T-cell count at HIV diagnosis which was lowest at the onset of the programme and increased over time, from 271 cells/mm3 to 340 cells/mm3 in the first 5 and the last 5 years, respectively. In addition, improved HIV prevention interventions might have played a role as well, such as increasing condom use, improved prevention of mother-to-child transmission coverage and an increasing number of HIV-positive individuals on ART, decreasing the risk of HIV transmission and leading to an increasingly healthy workforce.
Over a 15-year period, 14% of participants in the programme died. The mortality during the first year after ART initiation was stable at around 5% for all three cohorts. Although numbers and sample sizes are relatively small, this is remarkable, as, over time, ART was initiated at higher CD4+ T-cell levels (according to international treatment guidelines) and, therefore, participants were treated in an earlier stage of disease. Moreover, we found that the CD4+ T-cell count at diagnosis increased over time, suggesting that HIV-infected participants were identified earlier and could start treatment earlier. As it was hard to determine if a participant's death was HIV-related, it is also possible that the mortality rates as we found are partially due to other causes of death. This analysis was, however, beyond the scope of the current analysis.
The highest mortality rate was found among employees’ children of whom more than a quarter died during follow-up. These results are likely to be a reflection of the extra challenges of paediatric HIV treatment, such as the limited number of available antiretroviral drugs for paediatric usage and the fact that a child is dependent on its caregivers to receive and adhere to medication [6].
To our knowledge, this is the longest follow-up study evaluating the long-term outcomes of a workplace programme including HIV counselling, testing and treatment in sub-Saharan Africa. Relatively few studies have been published about the outcomes of private workplace programmes for HIV. A previous systematic review on workplace programmes for HIV and tuberculosis [7] identified reports of four workplace programmes in the private sector, including a previous evaluation of our current study [8]. Overall, these studies, conducted in Malawi, Zimbabwe and South Africa [9–11], found a higher HIV prevalence among their employees and a more advanced disease stage compared to our analysis. This could be explained by either a different, higher-risk target population in these studies or a better strategy to find those individuals most at risk to be screened for HIV.
Strengths and limitations
The major strength of this analysis is the programme's long implementation history of over 15 years, which is to our knowledge the longest documented follow-up period of a workplace programme in Africa. This analysis includes data from the earliest phase of HIV treatment in Africa in 2001 until the current situation, enabling us to evaluate the changes in treatment and treatment outcomes over time. Furthermore, our analysis includes not only employees of the company, but also their dependents, and retired employees. An important strength of this programme is that participants are entitled to lifelong ART provision, even if they are laid off work by the company, and do not need to transfer out to another ART clinic.
There are also some potential limitations to this study. First, as this was a retrospective analysis, not all data were available to study the desired outcomes. As a proxy for the success of the programme's interventions for HIV prevention, we would have liked to study the rate of mother-to-child transmission. However, because of privacy issues, data on employee's children and their mothers were separated, which made such an analysis impossible. In addition, as medical information and data from the human resource department were strictly separated, we did not have data on absenteeism or productivity to evaluate the impact of the workplace programme on the company's outcomes. In a previous analysis of this programme, however, it was reported that the costs of providing ART was estimated at around 45 US dollars per employee per year [5], which seems an acceptable price for a large international company. A study in South Africa showed that the workplace programme of a mining company was cost-effective, and that the costs per HIV-positive employee were reduced by 14% if ART was provided [12].
Second, all data were censored on 31 December 2015, which means that participants enrolled in the beginning of the programme had a longer follow-up than those enrolled near the time of censoring. In a longitudinal analysis this would create a bias, as early participants would have been more likely to have failed treatment or have died during follow-up. We therefore decided to conduct our analysis in three cohorts, providing only yearly estimates of proportions tested, on treatment and suppressed.
Finally, our outcomes might be subject to selection bias. We did an as-observed analysis, which means that we did not consider those participants who did not have viral load results or were lost to follow-up. It is possible that these specific participants had poorer treatment results than those who remained in care. In addition, providing HIV care through a workplace programme creates, by definition, some selection bias, as the employees of a company are generally healthier and have a higher socio-economic status compared to unemployed people. The excellent results as achieved in this programme, therefore, cannot be translated directly to the general population.
Future directions
In 2001, this programme was one of the first to provide free ART to its HIV-infected employees and their dependents in sub-Saharan Africa at a time when access to ART was still mainly limited to high-income countries. The fact that this company already in 2001 took the decision to embark on ART for its staff and dependents, supported by the PharmAccess Foundation, was a landmark event. This decision stimulated the development of other multinational HIV workplace programmes like those from CelTel, Diageo, Shell, Unilever, Anglo-American and KLM/Air France, as well as the Netherlands Embassies and the Netherlands Development Organization. From 2004 onwards, through the launch of the Global Fund, World Bank and PEPFAR programmes, the public sector in sub-Saharan Africa started extensive scaling-up of ART. The success of these programmes has been unprecedented. However, these programmes were not designed to encourage the creation of complementary public-private partnerships. Consequently, in many instances such programmes crowded out those ART programmes that were (on the brink of) being rolled out by the private sector. For example, the roll-out of a unique private sector risk equalization fund for HIV in Namibia that reached 15% of its formal sector workforce stagnated immediately after national provision of free ART by the government [13].
Our study shows that an employer ART programme, such as the Joep Lange HIV Workplace Program, is able to achieve excellent treatment outcomes at low costs (45 US dollars per employee per year [5]) that are reaching the treatment-related UNAIDS 90-90-90 targets, including in countries with some of the poorest HIV response indicators in the world, such as DRC, Republic of the Congo and Nigeria [14]. The continuous access to on-site quality care and antiretroviral drugs over the past 15 years, the limited investment needed for supplementary training of medical staff, procurement of some essential equipment, might be some of the reasons contributing to the success. The assistance of an independent NGO run by doctors with a background in HIV/AIDS treatment in Africa, focusing on continuous independent scrutiny of programme quality and costs, patient care and providing a double check on the legitimacy of expenditure has been important too. Government programmes might take these aspects, and in particular the integration of HIV care into general health care, as an example for optimizing HIV care in the public sector. This is particularly relevant in current times of stagnating or even decreasing (inter)national funding for HIV treatment. It is, however, important to note that the landscape of HIV care and treatment in sub-Saharan Africa has changed significantly over the past 15 years. While in 2001, there was hardly any alternative to the treatment provided by the Workplace Program, (free) public HIV care has now become the standard in most countries. A cost-benefit analysis of our programme in the current situation of both publicly and privately available HIV care could shed a light on the financial aspects of this programme. This analysis was, however, beyond the scope of our study.
In conclusion, this study gives insight in the developments of an HIV workplace programme in sub-Saharan Africa over 15 years. Antiretroviral treatment programmes, both public and private, require considerable and durable efforts to successfully reach the 90-90-90 targets. We show that it is possible to reach >90% on ART and >90% virological suppression in a private sector HIV programme in sub-Saharan Africa.
Footnotes
Acknowledgements
This study, conducted within the PharmAccess and Heineken HIV Workplace Program Group, is dedicated to late Professor Joep Lange, who established AIGHD and the PharmAccess Foundation and was the initiator and Principal Investigator of this ART programme starting as early as April 2001. Professor Lange's crucial contributions to making antiretroviral treatment available for Africans will always be remembered. The management of Heineken Int. is recognized for its timely and pioneering decision to embark on ART for all employees and dependents in times when ART for Africa was basically non-existent. This research was supported by a grant from Heineken International BV AMEE – 340110. PharmAccess and Heineken HIV Workplace Group Members: Vincent Janssens, Philippe Clevenbergh, Charles Kitenge, Richard Ajay, Gérard Ngendahimana Jean Pierre Kabarega, Laetitia Gahimbaza, Emmanuel Kamo, Paul Nsalou, Patrick Sianard, Ngozi Onya and Katinka van Cranenburgh.
This research was supported by a grant from Heineken International to AIGHD (AMEE – 340110).
SvdB and HR have been employed as (Chief) Medical Officer (CMO) by Heineken during part of the duration of this project. PC and HS remain employed by Heineken (CMO and Deputy CMO, respectively) and were employed by Heineken during part of the duration of this project.
The authors declare no competing interests.
